[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.

202]||ClickheretodownloadfreeAndroidapplicationfor

JSCIRES

RESEARCH ARTICLE

An index that rates relevance of scientific work in

biosciences: The scientific relevanceindex
Demosthenes B. Panagiotakos, Fragiskos G. Bersimis1
Department of Nutrition and Dietetics, School of Health Science and Education, Athens 17671, 1Department of Informatics and Telematics,
School of Digital Science, Harokopio University, Athens 17778, Greece

ABSTRACT
This work aims to present an objective index that rates scientific relevance(SR) of scientists published work, the SR
index, in a specific thematic field. The proposed index is calculated based on equally or weighted individual parameters that
measure scientists publication record and recognition, using easily accessible and unbiased data from existing bibliometric
databases. The application of the SRindex could be in any scientific field; an example is given here on cell and molecular
biology field.
Keywords: Scientific relevance, index, bibliometry

INTRODUCTION

To find the most relevant, but also the best, scientists in

order to develop a scientific committee for the evaluation,
and potential hiring, new faculty members for an Academic
Institution or Organization is a very difficult task. Similar to
the aforementioned mission, to rate and select the best new
faculty member among a group of potential candidates,
based on objective, quantitative criteria, is also very difficult.
During the past years, several academic measures have
been proposed, with the oldest ones being the number of
publications of an individual in a specific area and the total
number of citations of the published work that highlights
the recognition of the work. More recently, the conditional
number of citations after excluding selfcitations that
highlight the recognition of the work irrespective of the
individuals publication record(since it has been observed
*Address for correspondence: Prof. Demosthenes B. Panagiotakos
Email: dbpanag@hua.gr

Access this article online

Quick Response Code:
Website:
www.jscires.org
DOI:
10.4103/2320-0057.143688
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

that scientists with higher number of publications have

higher number of citations, too, because they usually cite
their own work) or focusing on a specific time period,
e.g.,the past 5years, have also been proposed and used.
There are also some more complicated indexes,[1,2] such
as the hindex which measures both the number of
publications and the number of citations per publication,
as well as the contemporary hindex which measures the
number of publications and the number of citations per
publication within a shorttime window,[3] the individual
hindex normalized by the average number of coauthors,[4]
the mindex(or mquotient) which has been defined as
hindex/the number of years since the first published
paper of the scientist,[5] the cindex that takes into account
not only the citations, but also the quality of the citations
in terms of the collaboration distance between citing and
cited authors, as well as many others with less use. Despite
the criticism the aforementioned indices have received,
ratings like these have been extensively used to measure the
impact of research and to justify tenure and in some cases,
funding decisions. At this point, it should be clearly defined
that bibliometric measures do not necessarily reflect the
quality of a published work. It could be argued that a
highquality paper, by not wellknown authors or/and
Institutions, published in a journal with limited promotion
resources, may receive less recognition than another work
of similar quality, but written by wellknown authors
15

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

and published in widely distributed or openaccess

journals. However, how quality of the publication can be
measured is a very difficult question, and it is not going
to be answered here.
Despite all the aforementioned considerations, the single
use of these measures to rate the most relevant, among
a set of scientists who have the apparent relevance in
a specific scientific area, shares several limitations and
biases. Thus, the purpose of this work was to propose
a multidimensional index(the scientific relevance[SR]
index) that rates the most relevant members among a set
of kscientists, based on their publication record, as well as
the international recognition of their scientific work. The
suggested index may be found useful for the selection of
optimal election committees in Academic Institutions or
Organizations, as well as for rating candidates for faculty
vacant positions.
METHODOLOGY

The SRindex is defined based on the following

parameters:(a) The total number of publications of the
iindividual, let Ni,(b) the number of the most relevant
publications to the specific scientific field, let v 1i ,[6,7] as
defined according to the exact match of the keywords that
characterizes the iindividual, using a keywords set of an
archiving database[Appendix], e.g.,Thompson
ReutersInstitute of Scientific Information, Scopus, (c)the
number of the less relevant publications, let v 2i , as defined
according to the keywords that characterizes the iindividual,
using a keywords set of the aforementioned archiving
databases and(d) the iindividuals hci index.
Let kbe the size of the set of all scientists according
to their apparent relevance in a scientific field. The
* of scientists is calculated for
maximum value max {}
i =1, 2,.., k
each of the following elements:
vi
1 i , which denotes the percentage of published work
N
that match in the exact scientific field(i.e.,all keywords
of individuals published papers are a superset of the
scientific topic)
vi
2 i , which denotes the percentage of scientific work
N
that belongs to a broader scientific field(i.e.,some of
the keywords of individuals published papers are a
superset of the scientific topic)
h i index of individuals published papers
hci index, which is the contemporary h i index for the
past m years.
16

The SRindex is then defined as follows:

v 2i
v 1i
Ni
Ni
SR - index =
+
vi
vi
max 1 i max 2 i
i =1, 2 ,.., k N
i =1, 2 ,.., k N
hci
hi
+
+
max h i
max hci
i =1, 2 ,.., k

{ }

i =1, 2 ,.., k

{ }

The SRindex is a continuous variable that takes values in

the interval[0, 4]. Greater values indicate

higher SR.
An issue that may arise is that SRindex does not fully
incorporate the true SR, neither the international
recognition nor the four parameters have the same
weighting. Therefore, a weighted SRindex is also proposed
here, where the weights b1,b2,b3, and b4 can incorporate,
depending on the values assigned, the true relevance and
recognition of an individuals scientific work. Weights
have to satisfy the equation: b1+b2+b3+b4=1 and should be
bj0, j = 1,2,3,4. The weighted SRindex is then defined
as follows:
v 1i
v 2i
Ni
Ni
Weighted SR - index = b1
+ b2
i
vi
v
max 1 i
max 2 i
i = 1, 2 ,.., k N
i = 1, 2 ,.., k N


i
hci
h
+ b3
+

b
4
max hci
max h i
i = 1, 2 ,.., k

{ }

i = 1, 2 ,.., k

{ }

The Weighted SRindex takes values in the interval[0, 1].

Higher values indicate greater SR among the members of
the potential candidates.
Both SRand weighted SRindex are increasing functions
of their components.
An Example
In this section, and as an example, it is attempted to rate
the SR of scientists published work using a data sample
from Greek academic registry in the scientific field
Molecular and cell biology. For this purpose, the SRindex
was calculated for each member of the aforementioned
sample, based on the number of scientists published papers
as appeared in Scopus until June 30, 2013, scientific
relevance was evaluated using keywords matching as
described in Table1, the full data are presented in Table2.
As mentioned, greater values of the SRindex rate the most
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

relevant scientists among the registrys members that are

initially considered as having the apparent relevance. For
example, the third members SRindex value was given by
the following formula, in which the publications number
was equal to 83, the most relevant publications number
was equal to 80, the wide relevant publications number was
equal to 81, the h 3index was equal to16 and the hc3index(for
the previous 5years) was equal to 8.

Table1: Parameters used to calculate the index of SR,

in the example presented in this work
Total number of
publications of
the ith individual
Number of
publications of
the ith individual
based on relevant
scientific work

Ni

Relevant topics

v1i

The following modules of the archiving

database Scopus, define the relevance
in this work(keywords based on Scopus)
Agricultural and Biological Sciences
Biochemistry, Genetics and Molecular
Biology
Chemistry

v 23
v 13
h3
N3
N3
SR - index 3 =
+
+
v i max h i
vi
max 1 i max 2 i i =1, 2 ,.., 20
i =1, 2 ,.., 20 N
i =1, 2 ,.., 20 N
hc3
0.96 0.98 16 8
+
=
+
+ +
i
0.98 0.99 36 10
max hc

Environmental Science

{ }

i =1, 2 ,.., 20

Health Professions
Immunology and Microbiology
Medicine
Neuroscience

{ }

Nursing
Pharmacology, Toxicology and
Pharmaceutics
Psychology

= 3.21 [ 0, 4 ]
The third members SRindex value was relatively high;
therefore, this scientist was classified in a high position
in the list. In addition, the weighted SRindex could be
calculated using weightings(subjective or objective). For
example, the value of the weighted SRindex for the third
member was given by the following formula, in which the
suggested weights b1,b2,b3, and b4 take the values: 0.35, 0.25,
0.20, and 0.20, respectively.

Social Sciences
Number of
publications of
the ith individual
based on wider
relevant scientific
work

v2i

Chemistry
Environmental Science
Biochemistry, Genetics and Molecular
Biology, Neuroscience
Medicine, Health Professions
Nursing, Pharmacology
Toxicology and Pharmaceutics

3
2

v
v 13
3
3
3
N
N
Weighted SR - index = 0.35
+ 0.25
i
vi
v
max 1 i
max 2 i
i = 1, 2 ,.., k N
i = 1, 2 ,.., k N


+ 0.20

hc3
h3
+
0
.
20

max h i
max hci

i = 1, 2 ,.., k

{ }

i = 1, 2 ,.., k

{ }

16
0.96
0.98
+ 0.25
+ 0.20
36
0.98
0.99
8
+ 0.20 = 0.84 [ 0, 1]
10

= 0.35

The weighted SRindex value was calculated for each

member of the data sample and is presented in the last
column of Table2. Depending on the values assigned
to the weights, evidently weighted SRindex value could
be altered in order to highlight the SR based on another
combination of weight values that better represent
Academic Institutions or Organization needs.
Regarding the mathematical properties of the SRindex, it
was an increasing function of all four components; with
more influential components being the hci index/max
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

The following modules of the archiving

database Scopus, define the wide
relevance in this work
Agricultural and Biological Sciences

Immunology and Microbiology

Psychology, Social Sciences
hindex of the ith
individual

hi

Number of papers with at least h citations

in the literature(including selfcitations)

Contemporary
hci
hindex of the ith
individual(5years)
SR=Scientific relevance

Number of papers with at least h citations

in the literature(including selfcitations)
within the past 5years

(hcindex), followed by the percentage of published work

that fully match in the scientific field and the hindex/max
(hindex).
DISCUSSION

In this paper, an index was proposed, which aimed to

rate most relevant scientists among a set of individuals
in a specific thematic module. This index may be a useful
tool for the selection of members, e.g.,of faculty election
committees for vacant positions in academic institutions,
or in various other research institutions, or for funding
purposes, in order to facilitate the work in finding the
most appropriate members for the evaluation committees.
This index may be also useful for the selection of the
17

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

Table2: Data used to calculate the SRindex in the example

Order

Number of
publications(N)

Full
relevance(v1)

91

87

214

205

0.96

210

83

80

0.96

81

v1
N
0.96

Wider
relevance(v2)
90

hindex

hcindex

SRindex

N
0.99

Weighted
SRindex

36

3.77

0.95

0.98

26

10

3.69

0.93

0.98

16

3.21

0.84

v2

73

71

0.97

72

0.99

20

3.14

0.83

41

40

0.98

40

0.98

23

2.92

0.78

110

55

0.50

98

0.89

18

10

2.91

0.70

53

31

0.58

51

0.96

30

2.90

0.72

60

59

0.98

59

0.98

18

2.79

0.76

79

76

0.96

77

0.97

15

2.78

0.75

11

54

42

0.78

53

0.98

18

2.68

0.70

12

63

40

0.63

60

0.95

28

2.69

0.68

13

38

36

0.95

37

0.97

15

2.66

0.73

14

26

24

0.92

25

0.96

2.63

0.72

15

74

37

0.50

70

0.95

13

2.63

0.65

16

28

21

0.75

26

0.93

17

2.57

0.68

17

69

30

0.43

67

0.97

27

2.57

0.63

18

106

37

0.35

95

0.90

25

2.56

0.61

19

61

30

0.49

55

0.90

17

2.38

0.60

20

59

45

0.76

55

0.93

14

2.31

0.62

SR=Scientific relevance

best candidates to cover a faculty or research position.

However, at this point, it should be strongly underlined
that the quality of a candidate, and not only the quantity,
even in terms of academic recognition, is what it matters.
Although it has been suggested that academic recognition
is a cornerstone of academic quality, this is not always
true, and various other characteristics also exist to define
quality. The advantage of the suggested index is that it
is not only unbiased, since it uses pure science metrics,
but it is also easily interpretable and applicable to a broad
range of scientific fields. The use of weights gives another
interpretation of SR values, based on what it is considered
more important, that is, the number of relevant publications
or the level of recognition or the contemporariness of the
published work. However, weights assignment should be
done with great care.
The proposed index has many advantages as well as
disadvantages. Amajor advantage is the fact that only
easily achievable through various scientific search
machines,[6,7] such as the Scopus and the ISI Thompson
Reuters archiving databases, as well as objective criteria
are used, whereas subjective assessment of the scientific
works relevance of the nominees is not involved, since
the keywords used may be a best approach to define what
is relevant and what is not. However, assessment of the
relevance may lead to distortions of the final outcome,
18

regarding the selection of the most appropriate members,

due to methodological, conceptual, and sometimes
emotional reasons. As regards the components of the
vi
vi
SRindex, the first two components(i.e., 1 and 2 ) rate
N
N
the publication impact of the individual, and are only
subject to the subjective determination of relevance, as
well as wider relevance. The next two components(i.e.,h i
and hci ) are considered as the most reputable evaluation
tools for the recognition of a scientists work. The hindex
attempts to measure both productivity and readability of
a scientific work of either a scientist or a group of
scientists or an academic department or university or even
a country.[1] The hindex was suggested by Jorge E. Hirsch,
physicist at University of California(San Diego, US), as a
tool for the determination of the relative quality among
theoretical physicists. Although the h(Hirsch) index is
generally accepted, there is a number of situations, in
which may provide misleading information. Specifically,
the index h does not take into account the number of
authors who participate in a paper. For solving this
problem, a corrected index has been proposed; the
individual hindex that normalizes, regarding the number
of coauthors, by dividing the hindex by the average
number of authors, who participate in a scientists work.
However, the position and the true contribution of an
individual in the writing group are disregarded and should
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

be further considered. It is also a fact that the hindex

occurs with large variability among different scientific
disciplines and sectors. For that reason, this index is not
recommended for readability comparisons among
scientists of different disciplines, such as medicine and
anthropology, as well as in sectors of the same discipline,
such as basic research and public health. In the case of
SRindex, this problem is, even partially, eliminated since
the sample space is created by scientists with relevant
scientific work(i.e.,keywords matching).
Besides, hindex is calculated from the total number of
publications of a scientist. This means that scientists
having a short career, regardless of the importance of
their scientific discoveries, are at an inherent disadvantage,
because they have not been exposed to the scientific
public sufficient time to receive citations of their
work.[1] However, as it has been noted by Hirsch, the index
proposed should be treated as a tool for the evaluation
of researchers at the same stage of their careers, and not
for historical comparisons. In this work, the hc index,
proposed by Sidiropoulos et al.,[8] was used in order to
smooth the previous problem. For this reason, the age
of the article, that is, the time that the article has been
exposed in the literature is taken into account. Finally, it
is generally accepted that hindex can be manipulated by
selfcitations. There are several academic debates being
done about whether selfcitations of a scientific paper
have to be taken into account or not. The selfcitations
inclusion may lead to the Matthew Effect[9] where the
rich in published work become richer in references
and the poor ones become poorer in citations of their
work. However, it may also be argued that large published
work is a denotation of scientists productivity, as well as
their continuity in time.

APPENDIX

Keywords based on Scopus archiving database for various

scientific fields.
AGRICULTURAL SCIENCES

The Agricultural Sciences category covers journals in

general agriculture, agricultural chemistry, and agronomy:
Agricultural engineering
Agronomy
Tillage research
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

Despite the majority of other scientometric indices

presented in the literature, the aim of the proposed index
was to become a tool, easy to use, and calculate that shows
the SR in a scientific area via an objective, as well as a fair
approach. The involvement of weights is an attempt to
highlight the absolute relevance and timeless readability
even more, but the inability to establish the weights
objectively may convert the index to a discriminatory
tool. Finally, the proposed index may hide some
potential interpretation problems in practice, but only its
implementation will highlight these methodological issues.
REFERENCES
1. HirschJE. An index to quantify an individuals scientific research
output. Proc Natl Acad Sci U S A 2005;102:1656972.
2. WendlMC. Hindex: However ranked, citations need context.
Nature 2007;449:403.
3. BatistaPD. Is it possible to compare researchers with different
scientific interests? Scientometrics 2006;68:17989.
4. von Bohlen Und HalbachO. How to judge a book by its cover?
How useful are bibliometric indices for the evaluation of scientific
quality or scientific productivity? Ann Anat 2011;193:1916.
5. Amors MB, DomingoFerrerJ, TorraV. Abibliometric index based
on the collaboration distance between cited and citing authors.
JInformetr 2011;5:24864.
6. Archiving and Data Base ISI Thompson Reuters. Available from:
http://www.thomsonreuters.com/. [Last accessed on 2014 Jun 30].
7. Archiving and data base Scopus. Available from: http://www.
scopus.com. [Last accessed on 2014 Jun 30].
8. SidiropoulosA, KatsarosD, ManolopoulosY. Generalized
Hirsch hindex for disclosing latent facts in citation networks.
Scientometrics 2007;72:25380.
9. MertonRK. The Matthew effect in science: The reward and
communication systems of science are considered. Science
1968;159:5663.
How to cite this article: Panagiotakos DB, Bersimis FG. An index
that rates relevance of scientific work in bio-sciences: The scientific
relevance-index. J Sci Res 2014;3:15-21.
Source of Support: Nil, Conflict of Interest: None declared

Agroforestry
Horticulture
Crop protection and science
Agrochemistry
Phytochemistry
Agricultural biochemistry
Food chemistry
Cereal chemistry
Carbohydrate and lipid research.
Food science and nutrition
Composition, additives, and contaminants
Microbiology and technology
19

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

Engineering and processing

Meat and dairy science
Nutrition science
Nutrition and metabolism
Nutritional biochemistry.
BIOLOGY AND BIOCHEMISTRY

Structure and chemistry of biological molecules

Molecular, cellular, and clinical studies of the endocrine
system
Regulation of cell, organ, and system functions by
hormones
Experimental research in general biology and biological
systems
Regulation of biological functions at the whole
organism level
Exploitation of living organisms or their components
Industrial microbiology
Pollution remediation
Industrial chemicals and enzymes
Biosensors
Bioelectronics
Pesticide development
Food, flavor, and fragrance industry applications
Waste treatment.
CHEMISTRY

Analytical chemistry
Spectroscopy
Instrumentation
Inorganic and nuclear chemistry
Organic chemistry
Physical chemistry
Polymer science
Food chemistry
Chemical methods and structures
Natural and laboratory syntheses
Isolation and analysis of clinically significant molecules
Medicinal chemistry
Chemical engineering.
CLINICAL MEDICINE

20

Anesthesia
Cardiovascular medicine
Dentistry
Dermatology
General and internal medicine

Endocrinology
Environmental medicine
Gastroenterology
Gynecology
Hepatology
Hematology
Nephrology
Nuclear medicine
Obstetrics
Oncology
Ophthalmology
Otolaryngology
Pediatrics
Pharmacology
Radiology
Toxicology
Respiratory medicine
Rheumatology
Surgery
Urology.
IMMUNOLOGY

Cellular and molecular studies in immunology

Clinical research in immunopathology
Infectious diseases
Autoimmunity and allergy
Hostpathogen interactions in infectious disease
Experimental therapeutic applications of
immunomodulating agents.
MICROBIOLOGY

Biology and biochemistry of microorganisms(bacterial,

viral, and parasitic)
Medical implications of the subsets of these organisms
known to cause diseases
Biotechnology applications of microorganisms for basic
science or clinical use.
MOLECULAR BIOLOGY AND GENETICS

Biochemistry
Molecular biology
Biophysics
Pharmacology
Receptor biology
Signal transduction
Regulation of gene expression
Developmental genetics and biology
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

[Downloadedfreefromhttp://www.jscires.orgonTuesday,February24,2015,IP:186.125.21.202]||ClickheretodownloadfreeAndroidapplicationforthisjournal
Panagiotakos and Bersimis: Scientific relevance index

Morphogenesis
Cellenvironment interactions
Molecular genetics
Mechanisms of mutagenesis
Structure, function, and regulation of genetic material
Clinical genetics, patterns of inheritance, genetic causes,
and screening and
Treatment of diseases.
NEUROSCIENCE AND BEHAVIOR

Cellular and molecular neuroscience

Neuronal development
Basic and clinical neurology
Psychopharmacology biobehavioral psychology
Molecular psychology
Neuronal function underlying higher cognitive
processes.
PHARMACOLOGY

Pharmacology
Pharmaceutics
Cellular and molecular pharmacology

Drug design and metabolism

Mechanisms of drug action
Drug delivery
Natural products
Xenobiotics
Mechanisms of action for clinical therapeutics.
Toxicology
Molecular and cellular effects of harmful substances
Environmental toxicology
Occupational exposure
Clinical toxicology.
PSYCHIATRY/PSYCHOLOGY

Biological
Clinical
Developmental
Educational
Mathematical
Organizational
Personal
Social
Diagnosis and treatment.

Announcement

Android App
A free application to browse and search the journals content is now available for Android based
mobiles and devices. The application provides Table of Contents of the latest issues, which
are stored on the device for future offline browsing. Internet connection is required to access the
back issues and search facility. The application is compatible with all the versions of Android. The
application can be downloaded from https://market.android.com/details?id=comm.app.medknow.
For suggestions and comments do write back to us.
J Scientometric Res. | JanApr 2014 | Vol 3 | Issue 1

21